This document discusses distributed systems concepts in both theory and practice. It covers operating systems research into concurrency and synchronization primitives. It then discusses distributed consensus algorithms from the 1970s and 1980s and the use of leases for distributed locking and leader election. It also discusses how staged event-driven architectures, leases, and allowing some inaccuracy can help build robust, scalable systems pipelines in industry. The document advocates using computer science research to inform design decisions and mitigate technical risks.